From Stealth to Seed Fund: Detecting Pre-Public Biotech Startups and Choosing Between NIH and NSF SBIR/STTR

TL;DR

• Stealth detection is a mosaic exercise, not a single signal. The strongest inferential indicators that a biotech/deep-tech company is still pre-public are: a filed-but-not-yet-published patent (a provisional, or a non-provisional inside the 18-month publication window), an SEC Form D or PitchBook/Crunchbase funding entry with no press release, incorporation records with no live website, "Stealth" LinkedIn titles, and simultaneous absence from NIH RePORTER and ClinicalTrials.gov. Cross-reference at least three signals before concluding a company is genuinely stealth.

• NIH vs. NSF comes down to whether your value proposition is a specific disease/health outcome or a broadly enabling platform. Choose NIH if your project is tied to a named disease, requires preclinical/human-subjects or FDA-regulated translation, and you can name the institute whose mission you serve. Choose NSF if your innovation is a high-risk, broadly applicable platform with strong intellectual and commercial merit and no clinical-study component (NSF explicitly returns clinical-study proposals without review).

• The dollars and mechanics differ materially in 2026. NIH's standard guideline caps (as of April 2026, per NIH SEED) are $323,090 (Phase I) and $2,153,927 (Phase II), with institute waiver ceilings up to $700,000 / $3,000,000; NSF Phase I is capped at $305,000 and Phase II at $1,250,000. NIH accepts unsolicited applications three times a year and lets you self-define the project; NSF requires an invited Project Pitch first and takes no clinical work. Both programs were reauthorized through September 30, 2031 by the Small Business Innovation and Economic Security Act (S. 3971 / P.L. 119-83), signed by President Trump on April 13, 2026.

PART 1 — HOW TO TELL IF A BIOTECH COMPANY IS STILL IN STEALTH

What "stealth" means and how long it lasts

Stealth mode is a deliberate strategy of minimal public exposure — no website, no press, minimal LinkedIn presence — while a company builds foundational science, IP, or go-to-market strategy. Analysts distinguish "total stealth" (no public-facing information at all; "Stealth Biotech" listed on LinkedIn) from "partial stealth" (a basic website or vague messaging exists but the core IP or business model is hidden). Deep-tech, AI, and biotech companies commonly run long-term stealth of one to five years because of extended R&D and regulatory timelines. The primary rationale is IP protection during the window before patents are secured, plus quiet talent recruitment and undistracted product development.

Signal 1 — Patent filing patterns (the most reliable inferable signal)

A provisional application is never published by the USPTO and automatically lapses at 12 months if not converted, so a company relying only on provisionals leaves no searchable patent footprint. A non-provisional (utility) application publishes 18 months after the earliest priority date under federal law — critically, that clock runs from the provisional filing date, so a converted application can publish sooner than founders expect. A company can therefore be operating with filed IP that is not yet public. Non-publication requests can also keep an application secret until it grants, but this binds the applicant to a US-only filing strategy.

Practical method: search USPTO Patent Public Search and Google Patents by inventor or founder name, not company name, since stealth entities often file under founders or assign the patent later. A named founder-inventor with no assignee, or an assignee shell with no web presence, is a strong stealth indicator.

Signal 2 — Funding records without public announcement

An SEC Form D must be filed within 15 days of the first sale of securities in a private raise, and it appears on EDGAR before any press coverage typically follows. However, research from Lehigh University finance professors documents that the majority of venture-capital-backed financing rounds are not accompanied by a Form D filing at all, and that rounds by firms with more proprietary information — especially early-stage, biotech, pharmaceutical, and high-tech companies — are less likely to file one. So a Form D is a positive signal when present, but its absence is not disconfirming for exactly the biotech companies you most want to find.

PitchBook and Crunchbase entries labeled "Stealth," "Stealth Mode," or "Stealth Biotech," or a round marked "Series Unknown" or undisclosed with no press release, are direct tells. Monitoring keyword filters on "stealth mode" in company descriptions can surface self-labeled companies even with free-tier database access.

Signal 3 — Incorporation vs. public disclosure gap

A live entity in a state's corporate registry (such as Delaware) with a reserved domain but no live website is a leading indicator. In the UK, a "dormant" status on Companies House can similarly signal a company quietly in development.

Signal 4 — Registry absences (the biotech-specific mosaic)

A genuinely stealth therapeutics or diagnostics company will typically be absent from all of the following: NIH RePORTER (no federal grants), ClinicalTrials.gov (no registered trials), SEC EDGAR (no public filings or Form D), FDA databases, and press or PR wires — while present in incorporation records and possibly a venture capital firm's portfolio. The combination of absences is the signal; any one absence alone is weak evidence.

Signal 5 — Team and LinkedIn hints

Founders listing "Stealth," "Stealth Startup," or "Building something new" alongside a real employment start date; senior scientific hires (VP of Research, Chief Scientific Officer) with no named employer; and job postings referencing NDAs all indicate an active but undisclosed company.

De-anonymizing tool — NIH RePORTER

The moment a stealth company accepts an NIH SBIR or STTR award, RePORTER exposes its legal name, principal investigator, project abstract, funding institute, and dollar amount. RePORTER's "similar projects" feature and activity-code filters (R41/R42 for STTR, R43/R44 for SBIR) let researchers map a company's technology, competitors, and assigned program officers. Several commercial lead-intelligence services now productize exactly this approach — mining SBIR.gov, NIH RePORTER, NSF Award Search, and SEC Form D data into daily "just-emerged" lead boards.

Checklist — is this company still in stealth? Score the mosaic across these indicators:

- No public website, or only a one-page holding page

- "Stealth" LinkedIn titles, or an unnamed employer for senior hires

- An active incorporation record with no product information

- Funding visible in PitchBook or Crunchbase but no press release, or a Form D on EDGAR with no announcement

- Provisional or unpublished (under 18 months old) patents filed under founder names, with no published applications

- Absent from NIH RePORTER

- Absent from ClinicalTrials.gov

- Present in a sector-focused VC's portfolio

Three or more of these present at once is high confidence the company is genuinely pre-public.

PART 2 — NIH SBIR/STTR: STRUCTURE, MONEY, CRITERIA, TRL

Structure

NIH is the dominant health-focused SBIR funder, awarding over $1.2 billion per year across its 27 institutes and centers. SBIR awards are coded R43 (Phase I) and R44 (Phase II or Fast-Track); STTR awards are coded R41 and R42. Applicants respond either to the Omnibus/parent solicitation or to targeted institute funding opportunity announcements, and can pursue Phase I, Fast-Track, or Direct-to-Phase-II pathways. The award always goes to the small business; STTR requires a subaward to a nonprofit research institution and permits a principal investigator who is not primarily employed by the company.

2026 funding amounts (per NIH's SEED program office, effective April 2026)

Phase I standard guideline (soft cap): $323,090 total costs, over six months to two years.

Phase II standard guideline: $2,153,927 total costs, over one to three years.

Commercialization Readiness Pilot: up to $4,191,495, over up to three years.

Institute waiver ceilings: up to $700,000 for Phase I and up to $3,000,000 for Phase II at institutes such as NINDS, NIDA, and NIDCD for SBA-approved waiver topics, plus a newer Phase IIB Strategic Breakthrough funding lane.

Budget guidelines are identical for SBIR and STTR, but each institute sets its own limits, so checking the funding document for your target institute can meaningfully change your available budget. Note that older sources may still cite the previous, now-superseded caps of $314,363 and $2,095,748.

Review criteria (2025 Simplified Framework)

Five regulatory criteria — Significance, Investigators, Innovation, Approach, and Environment — are now organized into three review factors: Factor 1, Importance of the Research (combining Significance and Innovation, scored 1 to 9); Factor 2, Rigor and Feasibility (Approach, scored 1 to 9); and Factor 3, Expertise and Resources (combining Investigators and Environment, evaluated as sufficient or insufficient). Scoring runs from 1 (exceptional) to 9 (poor), and the overall impact score is not a simple arithmetic average — a fatal flaw in the Approach criterion can sink an otherwise strong application. Approach is empirically the strongest predictor of whether an application gets funded. Study sections typically include 15 to 20 domain scientists, with about three reviewers assigned per application.

TRL expectations

NIH Phase I typically supports technology readiness levels 2 through 4 — exploratory research and feasibility work for medical technologies — aligning closely with the FDA approval pathway. Phase II moves toward prototype and validation work, roughly TRL 4 to 6. NIH study sections generally expect preliminary data even at Phase I, such as in vitro results, animal model data, or a published proof of concept; unlike some other agencies, topic responsiveness alone cannot compensate for a thin data package. Common rejection causes for biotech applicants include proposing Phase II-level scope within a Phase I application, and failing to address the FDA regulatory pathway in the commercialization plan.

Success rates

NIH is widely regarded as the most competitive SBIR program in the federal government. Phase I success rates have historically run roughly 15 to 18 percent, with institutes like NCI and NIAID being especially competitive; some more recent analyses cite an overall rate closer to 12 percent (roughly 10 percent for Phase I and 18 percent for Phase II). Resubmission materially improves odds — one historical NCI dataset showed roughly 14 percent success for original submissions versus roughly 24 percent for resubmissions.

PART 3 — NSF SBIR/STTR: STRUCTURE, MONEY, CRITERIA, TRL

Structure

NSF's SBIR/STTR program, branded "America's Seed Fund," awards about $200 million annually and funds roughly 400 companies per year, administered by the Directorate for Technology, Innovation and Partnerships. Awardees from fiscal years 2014 through 2023 raised an estimated $28 billion in private investment combined, with roughly 450 exits. NSF takes no equity and awardees retain full ownership of their intellectual property. The defining structural feature is the mandatory Project Pitch: applicants must submit a short pitch and receive an invitation from a program director before submitting a full proposal; uninvited proposals are returned without review. NSF funds broadly across nearly every technology area rather than soliciting specific topics.

2026 funding amounts (per NSF's May 2026 solicitation)

Phase I: up to $305,000, over six to eighteen months, inclusive of all direct and indirect costs, fee, technical assistance funding, and optional entrepreneurial training.

Phase II: up to $1,250,000, typically over 24 months.

Fast-Track pilot: up to $400,000 for Phase I plus up to $1,155,000 for Phase II.

Technical assistance: up to $6,500 in Phase I for commercialization activities, plus roughly $25,000 budgetable for entrepreneurial training programs.

Relevant life-science topic areas include Biological Technologies (covering synthetic biology and metabolic engineering, bioinstruments and biosensors, cell and tissue engineering, life-science research tools, microbiome work, and plant or animal biotechnology) and Biomedical Technologies (covering diagnostics, drug delivery methods, materials for biomedical applications, medical imaging, and monitoring devices).

Review criteria

NSF reviews proposals on three criteria: Intellectual Merit (the potential to advance knowledge via fundamental science or engineering that overcomes real technical risk), Broader Impacts (societal or economic benefit), and Commercial Impact or Commercialization Potential. Reviewers classify novelty into tiers, with a genuinely new scientific principle scoring highest and pure engineering optimization rarely scoring competitively. Proposals to NSF are kept confidential and do not constitute public disclosure — only a funded company's abstract eventually becomes public. Letters of support from customers are not allowed at the Phase I stage, but are required at Phase II; as of June 2026, these letters have returned as a requirement for NSF Phase II proposals.

TRL expectations

NSF Phase I typically starts around TRL 1 to 3 and advances to roughly TRL 3 to 4. The program is explicitly intended for high-risk research and development, not straightforward engineering or incremental product development. Crucially, clinical studies are considered non-compliant with NSF's scope — limited human-subjects work is allowed only for feasibility or proof-of-concept purposes, not clinical trials.

Success rates

NSF funds roughly 12 to 20 percent of Phase I applications historically, with Phase II rates considerably higher in some years. In one recent five-year window, 85 percent of Phase I awards went to companies with five or fewer employees, and 72 percent went to companies founded within the prior three years.

PART 4 — WHICH AGENCY FITS WHICH BIOTECH PROJECT

The core distinction is orientation: NIH is organized around a specific disease or health outcome and institute mission, while NSF is organized around broadly enabling platforms judged on intellectual and commercial merit.

On funding size, NIH's 2026 caps are higher — $323,090 for Phase I (up to $700,000 for waiver topics) and $2,153,927 for Phase II (up to $3,000,000 for waiver topics) — compared with NSF's $305,000 Phase I and $1,250,000 Phase II caps.

On timing, NIH Phase I projects run six months to two years, while NSF Phase I runs six to eighteen months.

On entry process, NIH accepts direct applications with no pre-invitation required, while NSF requires a mandatory invited Project Pitch before a full proposal can be submitted.

On deadlines, NIH has three standard cycles per year (roughly September, January, and April), while NSF ties deadlines to when a Project Pitch invitation is issued, with several windows across the year.

On review criteria, NIH scores Significance, Investigators, Innovation, Approach, and Environment on a 1-to-9 scale, while NSF scores Intellectual Merit, Broader Impacts, and Commercial Impact.

On clinical and human-subjects work, NIH actively supports it through Clinical Trial Optional funding announcements, while NSF does not allow it and returns such proposals without review.

On preliminary data, NIH generally expects it even at Phase I, while NSF is more tolerant of early-stage, high-risk concepts with less preliminary data.

On majority venture-capital ownership, both agencies allow it under specific conditions, with NIH having formally opted into this eligibility authority.

On typical technology readiness level at Phase I, NIH sits around TRL 2 to 4, while NSF sits around TRL 1 to 3.

Neither program takes equity; both are structured as grants.

Project-fit heuristics

Choose NIH if your product is a therapeutic, diagnostic, or device tied to a named disease or condition; if you need preclinical, animal-model, or human-subjects work; if your commercialization path runs through an FDA regulatory pathway such as an IND, 505(b)(2), 510(k), or PMA; if you can name the specific institute whose mission your work serves (searching RePORTER for similar prior awards can identify the right institute and program officer); if you want access to a larger Phase I or Phase II budget; or if your company is majority-owned by venture capital.

Choose NSF if your innovation is a broadly applicable platform or enabling tool — a research instrument, biosensor, synthetic-biology chassis, computational biology tool, or manufacturing and biomanufacturing process — whose value isn't tied to one specific disease; if you are pre-preliminary-data and carrying high technical risk; if there is no clinical-study component to your work; and if you can clearly articulate both fundamental technical risk (intellectual merit) and a genuine commercial market.

A note on pursuing both agencies: it is illegal to accept duplicate funding for the same work, and both agencies require disclosure of overlapping or equivalent proposals on Current and Pending Support forms. Program officers at NIH and NSF do coordinate on overlap. You may pursue genuinely distinct projects at each agency — for example, an NSF-funded enabling platform alongside an NIH-funded disease-specific application — but the scope of each must be clearly delineated. NIH will not review duplicate or highly overlapping applications simultaneously, even across different activity codes, allows one resubmission within a 37-month window, and HHS now caps submissions at nine per small business per fiscal year. NSF allows only one proposal per principal investigator under review at a time, and permits resubmission of a returned-without-review proposal under the same Project Pitch within two subsequent deadlines.

CASE EXAMPLES

NSF — platform and enabling technology fit: Shasqi, a company developing targeted "click chemistry" drug activation led by a physician-chemist founder, began with an NSF SBIR grant when it was too early-stage for venture capital. Yesse Technologies, developer of a "nose on a chip" odor-sensing platform, progressed from a $225,000 Phase I award to a $750,000 Phase II award. Azitra, which engineers the skin microbiome to treat skin disease, and Caption Health, which built an AI-guided ultrasound platform, are both NSF-funded life-science companies. These companies share the broad-platform, high-merit profile that NSF tends to reward.

NIH — disease-specific translation fit: An economic-impact study of NCI's SBIR program found that 444 companies received NCI Phase II funding across 690 projects between fiscal years 1998 and 2010, against an NCI investment of $787 million; 53 percent of those projects resulted in product or service sales, generating an estimated $9.1 billion in total sales and roughly 108,000 jobs. As a representative example, one company received a $1.5 million Phase II award from NINDS in 2018 to develop a bioabsorbable surgical clip, using the funding to accelerate development and cover commercialization costs. Therapeutics and diagnostics tied to a specific condition — cancer, neurological disease, infectious disease — represent the NIH sweet spot.

RECENT POLICY CHANGES EVERY BIOTECH APPLICANT SHOULD KNOW (2025-2026)

Reauthorization through 2031: After the prior authorization lapsed on September 30, 2025 — halting new awards during roughly a six-month freeze — the Small Business Innovation and Economic Security Act passed the Senate by voice vote on March 3, 2026, passed the House 345 to 41 on March 17, 2026, and was signed into law by President Trump on April 13, 2026, extending SBIR/STTR through September 30, 2031. The roughly $6 billion measure also created a new "strategic breakthrough award" of up to $30 million, requiring a 100 percent private-capital match.

New national-security and foreign-influence rules: National-security reviews are now mandatory, and awards are prohibited to companies with specified foreign ties. HHS due diligence now covers cybersecurity, patent analysis, employee analysis, foreign ownership and affiliations, investment relationships, licensing, joint ventures, and business relationships with countries of concern. Applicants must be prepared to document ownership structure, capital origin, IP control, and any foreign affiliations of key talent.

NIH-specific changes: HHS now limits SBIR/STTR submissions to nine per small business per fiscal year, and small-business applications are no longer eligible for NIH's late-submission policy. NIH reissued its parent funding announcements in late May 2026, with the submission window reopening August 5, 2026, and the first standard deadline around September 5 or 8, 2026, followed by January 5, 2027, and April 5, 2027. NIH maintained its study-section infrastructure through the funding freeze, allowing a fast restart.

NSF-specific changes: NSF released a new solicitation in May 2026, reframed around developing deep technologies that advance U.S. competitiveness and security, replacing the prior solicitation series. New Project Pitch submissions opened June 2, 2026, with full-proposal deadlines including late July 2026 and early November 2026, then settling into a recurring pattern of the first Wednesday in November, first Thursday in March, and first Wednesday in July annually. Letters of support also returned as a requirement for NSF Phase II proposals as of June 2026.

Budget-environment caveat: The administration's FY2026 discretionary budget request proposed a roughly 37 percent cut to NIH and more than a 50 percent cut to NSF. Because each agency's SBIR/STTR set-aside is a fixed percentage of extramural R&D spending under the 2026 reauthorization, enacted cuts of that scale would proportionally shrink these programs. As of this writing, these are proposals, not enacted appropriations, and should be treated as a risk factor rather than a certainty.

RECOMMENDATIONS

For founders deciding between NIH and NSF, a staged approach works well:

First, classify your value proposition. Write one sentence: "Our technology enables or treats ___." If the blank is a named disease or an FDA-regulated product, start with NIH. If it's a broadly applicable capability, start with NSF. If any Phase I aim involves a clinical study, NSF is disqualified outright — go to NIH.

Second, for NIH candidates, run the RePORTER test. Search NIH RePORTER for prior R43/R44 awards matching your keywords. If several overlapping projects were funded by one institute, that institute is your target, and it's worth contacting its SBIR program officer at least a month before the deadline.

Third, for NSF candidates, run the Project Pitch test early. Draft and submit the pitch as soon as possible; the invitation — or its absence — is a fast, free signal of fit. Without an invitation, a full proposal isn't worth the effort.

Fourth, match your technology readiness level and data package to the agency. If you have preliminary in vitro or animal data and a clear regulatory story, NIH rewards it. If you're pre-data with high technical risk and a genuinely fundamental science question, NSF is the better home.

Fifth, optimize the budget ask. NIH candidates should check whether their target institute has waiver topics allowing requests up to $700,000 for Phase I or $3,000,000 for Phase II without a separate individual waiver. NSF candidates should budget the $6,500 in technical assistance funding and roughly $25,000 in entrepreneurial training within the $305,000 cap.

Sixth, consider a deliberate dual-agency strategy for more advanced situations — funding an enabling platform at NSF and a specific disease application at NIH — but delineate the scope of each precisely and disclose both on Current and Pending Support forms to avoid overlap issues.

For consultants qualifying stealth-company prospects, build a mosaic score using the eight-point checklist above, treating any single signal as weak evidence and requiring at least three simultaneous signals for a "genuinely stealth" classification. Automate a primary-source sweep across SEC EDGAR Form D filings, USPTO and Google Patents searches by founder name, state incorporation records, NIH RePORTER, NSF Award Search, and ClinicalTrials.gov. The transition from stealth to public visibility is often first detectable through a RePORTER award or a Form D filing — sometimes days or weeks before any press coverage follows. Time outreach to that emergence moment: a newly posted SBIR award, a first published patent, or a first Form D filing marks the window when a stealth biotech becomes a warm prospect for funding-strategy help.

Two benchmarks worth watching: if enacted FY2026 or FY2027 appropriations cut NIH or NSF R&D funding substantially, expect paylines to tighten and success rates to fall below the historical 15-to-18-percent (NIH) and 12-to-20-percent (NSF) ranges, which would argue for prioritizing resubmission quality and earlier-cycle submissions. And if a company becomes majority-owned by a single VC or private equity firm at more than 50 percent, it loses SBIR eligibility at most agencies — restructuring the cap table so no single investor exceeds 50 percent ownership should happen before applying.

CAVEATS

Cap figures move over time. The most current NIH guideline caps as described here — $323,090 for Phase I and $2,153,927 for Phase II — reflect NIH's SEED program office guidance as of April 2026; older sources still cite the previous $314,363 and $2,095,748 figures, or direct-cost-only figures that understate the true total. Always confirm against the live solicitation before finalizing a budget.

Success-rate figures vary by source and year, and are sometimes conflated across grants versus contracts, or Phase I versus Phase II. Treat the ranges given here as directional rather than precise.

Stealth detection is inferential by nature. Absence from a registry can reflect genuine stealth, or simply reflect a company that hasn't yet needed federal funding or registered a trial. False positives are common, and no single signal should be treated as proof — notably, the biotech firms most worth finding are also the least likely to file a Form D.

Policy is in flux following the recent reauthorization. New foreign-influence and national-security due-diligence rules, the nine-submission HHS cap, and reissued funding-announcement numbers are all recent developments — verify current deadlines and requirements directly against the agencies before relying on them.

Finally, the FY2026 budget cuts described above are proposals, not enacted law. Any statement about future NIH or NSF SBIR funding shrinkage is contingent on appropriations that had not been enacted as of this writing.